Hg, Cu, Pb, Zn, Ni - American Chemical Society

Apr 10, 2015 - as well as the top and bottom clamp, were machined in-house from Teflon ..... Jonathan Newland for the photograph in Figure 1c and for...
0 downloads 0 Views 1MB Size
Subscriber access provided by University Libraries, University of Memphis

Article

Direct Identification and Analysis of Heavy Metals in Solution (Hg, Cu, Pb, Zn, Ni) using In-Situ Electrochemical X-Ray Fluorescence Glen D. O'Neil, Mark E. Newton, and Julie V. Macpherson Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b00597 • Publication Date (Web): 10 Apr 2015 Downloaded from http://pubs.acs.org on April 14, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Direct Identification and Analysis of Heavy Metals in Solution (Hg, Cu, Pb, Zn, Ni) using In-Situ Electrochemical X-Ray Fluorescence Glen D. O’Neil,+ Mark E. Newton§ and Julie V. Macpherson+* Department of Chemistry+ and Department of Physics,§ University of Warwick, Coventry, CV4 7AL The development and application of a new methodology, in-situ electrochemical X-ray fluorescence (EC-XRF), is described which enables the direct identification and quantification of heavy metals in solution. A freestanding film of boron doped diamond serves as both an X-ray window and the electrode material. The electrode is biased at a suitable driving potential to electroplate metals from solution onto the electrode surface. Simultaneously, X-rays which pass through the back-side of the electrode interrogate the time-dependent electrodeposition process by virtue of the XRF signals, which are unique to each metal. In this way it is possible to unambiguously identify which metals are in solution and relate the XRF signal intensity to a concentration of metal species in solution. To increase detection sensitivity and reduce detection times, solution is flown over the electrode surface using a wall-jet configuration. Initial studies focused on the in-situ detection of Pb2+ where concentration detection limits of 99 nM were established in this proof-of-concept study (although significantly lower are anticipated with system refinement). This is more than three orders of magnitude lower than that achievable using XRF alone in a flowing solution (0.68 mM). In-situ EC-XRF measurements were also carried out on a multi-metal solution containing Hg2+, Pb2+, Cu2+, Ni2+, Zn2+ and Fe3+ (all at 10 µM concentration). Identification of five metals was possible in one simple measurement. In contrast, whilst anodic stripping voltammetry (ASV) also revealed five peaks, peak identification was not straightforward, requiring further experiments and prior knowledge of the metals in solution. Time-dependent EC-XRF nucleation data for the five metals, recorded simultaneously, demonstrated deposition rates were similar. Studies are now underway to lower detection limits and provide a quantitative understanding of EC-XRF responses in real, multi-metal solutions. Finally, the production of custom designed portable in-situ EC-XRF instrumentation will make heavy metal analysis at the source a very realistic possibility. Submitted to Analytical Chemistry, February 2015; revised April 2015

1 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

INTRODUCTION Heavy metal contamination caused by increased industrialization is a well-documented, serious problem due to the negative effects these chemical species have on human health, safety and the environment.1 Heavy metals are not biodegradable and can accumulate in vital organs such as the liver and kidneys, as well as the central nervous system.2,3 As a result, long-term exposure to even trace levels can be potentially disastrous. This is further exacerbated by considering humans are at the top of the food chain, and ingesting foodstuffs contaminated with heavy metals will have the same effects as direct exposure. For these reasons, world governments now place strict guidelines on acceptable levels.4,5 Many techniques exist for identifying and quantifying trace levels of heavy metals, including inductively coupled plasma mass spectrometry,6 atomic absorption and emission7 and colorimetric methods based on host-guest chemistry.8,9 These methods can be high cost, require complex sample or substrate preparation and are often non-portable, which limits their use to the laboratory. Portable, low cost electrochemical methods, such as anodic stripping voltammetry (ASV)10,11 and potentiometry using ion-selective electrodes (ISEs)12,13 have showed promise for trace analysis in-situ. However, these electrochemical techniques often suffer when placed in “real” unknown multi-metal solutions, due to the complexities of stripping peak interpretation (for ASV),14,15 and

co-extraction (for ISEs),12,16

making

absolute quantification very difficult. Energy dispersive (ED)-X-ray fluorescence spectroscopy (XRF), which enables chemical identification of individual elements present in a sample, is commonly used for the detection of heavy metals, typically at the ppm level, but is mostly employed with solid samples and powders, due to experimental set-up.17–19 Increased detection sensitivity can be attained through the use of total reflection XRF, which reduces background scatter.20,21 Direct XRF analysis of liquid (aqueous) samples is possible, but detection limits are severely limited (vide

2 ACS Paragon Plus Environment

Page 3 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

infra).22 In order to increase detection sensitivity, the solution is often evaporated onto a suitable carrier, leaving the salt for analysis.23 Recently, we introduced electrochemical X-ray fluorescence (EC-XRF), as a means to chemical identify trace metals in solution at sub-ppb levels.24 Electrochemical deposition was used to pre-concentrate heavy metals from solution onto a freestanding polycrystalline borondoped diamond electrode (BDD), which serves as an excellent EC-XRF substrate due to its wide solvent window, low background currents,25 and low background fluorescence of carbon and boron. However, the XRF measurements took place ex-situ with the electrode removed from solution and analyzed on the bench-top. To address the issue of portability and measurement at the source we present in-situ ECXRF methodology, which enables simultaneous electrodeposition of metals on the electrode surface coupled with XRF analysis, paving the way for in-situ identification and analysis of heavy metals, using this technology. Here, solution is delivered to the BDD thin (~ 250 µm) window electrode using a wall-jet type electrode configuration, to increase detection sensitivities and decrease analysis times compared to stationary solutions. Preconcentration of the metal takes place on the upper electrode surface whilst the high energy X-rays used to excite the fluorescence pass through from the back face. The secondary “fluorescent” X-rays emitted by the electrodeposited metal pass through the BDD to the detector positioned below, as shown in Figure 1a. The use of thin BDD window electrodes reduces background X-ray scatter from the electrode matrix. Using in-situ EC-XRF we demonstrate that both single (Pb2+) and complex (Hg2+, Cu2+, Pb2+, Ni2+, Zn2+) mixtures of heavy metals can be unambiguously identified in solution at low concentrations, suggesting that field-based measurements in real environments is a future possibility. In this first demonstration detection sensitivities of ~ 10’s ppb are reported but further increases are achievable with system refinement. The approach also interestingly,

3 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

enables real time information to be acquired on metal deposition, which is likely to be of interest in understanding electrodeposition characteristics especially in multiple metal environments.

EXPERIMENTAL SECTION Materials: Hexaamineruthenium(III) chloride ([Ru(NH3)6]Cl3) (Acros Organics); lead (II) nitrate (Pb(NO3)2; 99.999%, trace metal basis), copper (II) nitrate (Cu(NO3)2; 99.999%, trace metal basis), nickel (II) nitrate hexahydrate (Ni(NO3)2·(H2O)6; 99.999%), zinc (II) nitrate hexahydrate (Zn(NO3)2·(H2O)6; 98%), iron(III) chloride (FeCl3; 99.99%, trace metal basis), and potassium nitrate (KNO3), all from Sigma-Aldrich, were used as received. Mercury (II) nitrate monohydrate (Hg(NO3)2·H2O; >98.5%) was from Merck. All solutions were prepared using 18.2 MΩ cm-1 deionized water (Millipore). Solution pH was measured using a MetlerToledo glass-membrane pH electrode. Freestanding, conducting polycrystalline BDD wafers with boron concentrations ~ 3 × 1020 cm-3 and polished to 250 µm thickness, with a rms surface roughness < 5 nm, were supplied by Element Six Ltd. (Harwell, UK). The wafers were laser machined into 25 mm diameter (working) and 16 mm diameter (counter) electrodes using laser micromachining methodologies described previously.26 Prior to use, BDD electrodes were acid cleaned by boiling in concentrated H2SO4 (98%), supersaturated with KNO3 in order to remove any non-diamond-carbon (NDC) from the laser micromachining process.27 BDD 1 mm diameter glass sealed macroelectrodes were prepared as described previously.26,28 In-situ EC-XRF apparatus and operation: The in-situ EC-XRF electrochemical cell is based upon a wall-jet configuration, in which an impinging jet of solution flows axially onto the electrode surface and, upon contact, radially across the electrode surface.29,30 The structural components of the cell, which defined the positions of the inlets, outlets, working, 4 ACS Paragon Plus Environment

Page 4 of 22

Page 5 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

counter and reference electrodes, as well as the top and bottom clamp, were machined inhouse from Teflon according to the drawing shown in Figure 1b. The design contains a single 500 µm diameter inlet which is placed in the center of Teflon block. In order to maintain uniform hydrodynamics within the cell, the single inlet empties into six outlets (1.6 mm ID) which are evenly spaced in the cell at positions in line with the circumference of the working electrode. In order to maintain an even pressure drop between the inlet and outlets, all six sections of outlet tubing were epoxied into a custom manifold, to form a single waste outlet. All of the tubing was fixed in place using Araldite 5 minute epoxy. Fluid flow was controlled using an HPLC pump (Gilson Model 305, France) flowing through a manometric module to decrease noise (Gilson model 806, France). In order to form an electrochemical cell, the Teflon body was placed directly on top of the BDD electrode, and assembled using a clamp and M4 screws (Fig 1b). A 2.0 cm diameter Oring was employed to create a leak-proof seal and determine the diameter of the working electrode. The channel height was fixed to be >200 µm (actual height measured using interferometry is 650 ± 10 µm) such that the jet of solution was well-established before impinging onto the electrode surface.30 Two counter electrode configurations were initially tested. The first consisted of three Pt wire coils (250 µm; Goodfellow) electrically connected in parallel, which were inserted into three of the six outlets. In the second, the counter was laser machined from 350 µm thick freestanding conducting BDD in a shape that matched the underside of the Teflon body (including the inlet hole). A 3 mm diameter hexagonal array of 60 µm diameter holes was positioned around each of the six outlets, to increase electrode surface area in order to obtain comparable working and counter electrode areas. The counter was then epoxied (Epoxy Resin RX771C/NC, Aradur Hardener HY1300GB, Robnor Resins) onto the upper wall of the Teflon flow channel so the front face was co-planar with the surrounding Teflon, as shown in Fig 1b.

5 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Ohmic contact to the BDD counter was made by graphitizing a small section of the rear of the electrode, using laser micromachiner and contacting a Cu wire using AgDAG (Silver conductive paint, RS, UK).31 Ohmic contact to the BDD working electrode was made by sputtering a thin Ti/Au (10 nm and 300 nm, respectively) contact onto the backside of the BDD surface, and subsequently annealing at 400 oC for 5 hours. A thin ring of Cu foil (2.5 cm outer diameter and 2.4 cm outer, 100 µm thick) with a 1 cm ‘tongue’ (Fig. 1b) was cut and placed between the backside of the BDD disk and the Teflon clamp (Fig. 1b) and was used to provide an accessible electrical contact. In all experiments, Ag/AgCl wire placed in one of the outlets was used as a pseudo-reference electrode. Electrochemical deposition: All electrochemical experiments were performed using either an Ivium CompactStat (Alvatek Ltd., U.K.) or a CHI 730A (Austin, TX, USA) potentiostat controlled using a PC. Prior to deposition the BDD electrode(s) were cleaned with 0.05 µm alumina slurry (Micropolish, Buehler, Germany) using a microcloth polishing pad or a disposable cotton bud. Background XRF spectra of the BDD electrode, in air, were collected after cleaning to ensure the surface was metal free. The metal electrodeposition parameters (deposition potential, flow rate, deposition time) were initially optimized to maximize the XRF signal intensity for Pb2+ electrodeposition in 0.2 M KNO3 (pH ~ 5.5) (see Electronic Supporting Information ESI, Section 1), which served as the model heavy metal in these proof-of-concept studies. X-Ray fluorescence. In-situ EC-XRF was performed using a commercial ED-XRF (NEX CG, Rigaku, Japan). An X-ray tube containing a Pd anode was run at 50 kV with a 1 mA current with a Mo secondary target (Kα = 17.48 keV). The inlet, outlet, and electrical leads were fed into the instrument through the vacuum inlet at the rear of the vacuum chamber (Figure 1c). In this way the X-ray chamber remained fully sealed during in-situ ECXRF measurements to contain stray X-rays. All spectra were collected under ambient

6 ACS Paragon Plus Environment

Page 6 of 22

Page 7 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

conditions. The sampling time varied depending on the demands of the experiment, and is described in detail alongside the results. The interrogation area of the X-ray source was elliptical being ~1.2 cm in the widest part, decreasing to ~1.0 cm in the narrowest. The walljet electrode was placed over the X-ray optics, so the X-rays entered from the backside of the electrode. The Cu foil connection faced towards the source. All XRF spectra were collected with the device fully assembled within the X-ray chamber in the configuration shown in Fig. 1c.

Figure 1: Experimental configuration of the in-situ EC-XRF device. (a) Schematic of the wall-jet EC-XRF cell and mode of operation; (b) schematic of the device assembly; (c) photograph of the wall-jet device inside the X-ray chamber. RESULTS AND DISCUSSION Device characterization: A wall-jet configuration was employed as a means of increasing mass transport to the electrode during electrodeposition. In the set-up considered, solution

7 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 22

axially impinges on the electrode and after contact flows radially across the surface.30,32 As the nozzle diameter is smaller than the electrode diameter, the highest fluid velocities occur near the center of the electrode, and steadily decay towards the edges. This electrode configuration allows for well-defined and variable mass transport rates, which are necessary for reproducible electrochemical depositions during the preconcentration stage of the experiment. The electrochemical response of the wall-jet electrode has been well-described in the literature,32,33 where the limiting current is given by equation (1).33 ଷ/ସ ∗

݅௟௜௠ = 1.597݊‫݇ܨ‬௖ ܴ ଷ/ସ ߭ ିହ/ଵଶ ‫ ܦ‬ଶ/ଷ ܽିଵ/ଶ ܸ௙

ܿ

(1)

n is the number of electrons transferred, F is Faraday’s constant (96485 C mol-1), kc takes a value between 0.86 and 0.90, R is the electrode radius = 1.0 cm, v is the kinematic viscosity, D is the diffusion coefficient, a is the diameter of the inlet = 0.05 cm, Vf is the volume flow rate and c* is the bulk concentration. Equation (1) assumes laminar flow across the surface of the electrode and that the wall-jet is not impeded by the top of the channel. The inset to Figure 2a shows chronoamperometric i-t curves, recorded over 15 s, for the reduction of 0.5 mM Ru(NH3)63+ in 0.2 M KNO3 for Vf in the range 0 – 18 mL min-1. The potential was stepped from 0 V vs. Ag|AgCl where no faradaic reaction occurs to a potential where the current response is mass-transport limited (-0.5 V vs. Ag|AgCl). At all but the slowest Vf does the current attain a steady state, with the time taken to reach ilim decreasing as Vf increases. A plot of ilim (n = 3) versus Vf¾ is shown in Figure 2a with a linear regression showing excellent agreement (R2 = 0.9999) between ilim and Vf3/4 at flow rates above 10 mL min-1. At the lowest Vf the current signal starts to deviate from the theoretical response, which is characteristic of diffusion starting to dominate over convection. The gradient of the linear regression line (4.27 (±0.03) × 10-5 mA min3/4 mL-3/4) agrees very well with that calculated

8 ACS Paragon Plus Environment

Page 9 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(4.1× 10-5 mA min3/4 mL-3/4) assuming kc = 0.9, R = 1.0 cm, v = 8.8 × 10-3 cm2 s-1, D = 8.8 × 10-6 cm2 s-1,34 a = 0.05 cm, c* = 0.5 × 10-6 mol cm-3and n = 1.

Figure 2: Characterization of the in-situ EC-XRF flow cell. (a) ilim versus Vf3/4 (•) for the diffusion-limited oxidation of 0.5 mM Ru(NH3)63+ in 0.2 M KNO3 and comparison with theory (___), equation 1. Inset: Amperometric i-t curves over the Vf range 2 - 18 ml min-1. Error bars represent one standard deviation from three replicate measurements. (b) XRF background spectra collected with either a platinum (red line) or BDD (black line) counter electrode positioned within the EC-XRF cell. (c) Quantifying detection limits for XRF-only 9 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

measurements of [Pb2+] in the flow cell. Plot of XRF peak height, measured at 10.56 keV, versus [Pb2+] over the range 1 – 50 mM at Vf = 14 ml min-1. The spectral properties of the electrochemical cell were investigated with the cell simply filled with solution (10 µM Pb2+) and placed over the X-ray window. Figure 2b shows two XRF spectra over the energy range 6-15 keV (vide infra). The black and red lines show the response when employing a Pt and BDD counter electrode respectively. In both cases, there is no evidence of a PbLα line at 10.56 keV, indicating that a concentration of 10 µM Pb2+ is well below the limit of detection for Pb2+ in this configuration. Both counter electrode arrangements show small peaks for FeKα and ZnKβ/Kα. The former is due to Fe present in the stainless steel X-ray chamber, whilst the Zn peak originates from Zn present in the O-ring used in these studies (validated using XRF experiments not detailed here). The cell containing the Pt counter electrode shows three distinct peaks at 9.4 keV, 11.1 keV and 12.9 keV corresponding to the PtLα, PtLβ, and PtLγ lines. Note, the intensity of PtLα masks the ZnKβ (9.59 keV) response. Use of a BDD counter obviously removes the Pt XRF spectral lines, and hence was used for all further studies. XRF spectra (from 5.5 to 14 keV), were recorded with solution flowing through the cell at Vf = 14 ml min-1 (Figure 2c) for different Pb2+ concentrations (1 – 50 mM). The response of the PbLα peak intensity correlates well with [Pb2+] over the concentration range investigated (R2 = 0.996), with a detection limit (LOD = 3σ; where σ is the standard deviation of the noise) of 0.682 mM. This result establishes a benchmark for XRF-only measurements in flowing solutions (for our setup) and is consistent with previous work where XRF has been used previously to measure metal ion concentrations in solution.22 This result also highlights that some form of preconcentration is necessary in order to significantly reduce detection sensitivity when making XRF measurements directly in solution. In-situ EC-XRF: For temporal measurements it is important to understand the effect of the rate of spectral acquisition on the XRF response. This was investigated using flowing 10 ACS Paragon Plus Environment

Page 10 of 22

Page 11 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

solution (Vf = 14 mL min-1) containing background electrolyte only (0.2 M KNO3). Figure S2 in ESI 2 shows histograms of the peak area recorded over the range 10.1-11 keV, where the PbLα peak would reside if Pb2+ were present in solution. This represents the noise level of the blank signal. Spectra were acquired over one hour using collection times of 1, 10, 30 and 50 seconds. As is clear from Figure S2, as collection time decreases, the FWHM of the “noise” data increases, indicating lower signal-to-noise ratios. As a result for all the following experiments, an X-ray collection time of 50 seconds was chosen to maximize detection sensitivity, whilst still maintaining a useful (< 1 min) temporal resolution. To investigate electrochemical preconcentration and showcase the advantages of performing simultaneous electrochemical deposition and temporal analysis in-situ using ECXRF, a solution containing 10 µM Pb(NO3)2 was analysed. A Vf = 14 mL min-1 and deposition potential, Edep, of -1.5 V vs. Ag|AgCl were employed (ESI Section 1). Figure 3a displays typical XRF data, over the energy window 9.75 – 11.5 keV (PbLα = 10.56 keV) collected at 5 min intervals prior to and during electrochemical deposition (tdep = 60 min). The corresponding i-t curve recorded during electrochemical deposition is shown in the inset to Figure 3b. The red traces in Figure 3a show three background spectra recorded with no potential bias applied to the electrode. As expected for a 10 µM solution of Pb2+ no XRF signature for Pb2+ is evident in the region of the PbLα line. However, under potential bias (purple and blue lines), not only is a signature observed for PbLα but the PbLα line grows in intensity with time, due to an increase in the amount of electrodeposited Pb on the electrode surface. The amount of charge passed during electrodeposition was calculated from the i-t response (shown in the inset to Figure 3b) as 51.4 mC (over 60 mins) which corresponds to 2.7 × 10-7 moles of Pb, assuming the current passed is due only to the electroreduction of Pb. This would equate to a

11 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

maximum volume of 4.93 ×10-6 cm3 of Pb deposited on the surface (assuming a density of 11.34 g cm-3). To visualize the intensity change in more detail, the acquired spectra were integrated and the resultant peak area plotted as a function of t as shown in Figure 3b. The plot on the time (x) axis has been shifted so that electrochemical deposition starts at t = 0 s. Here it can be seen that Pb electrochemical deposition, over this time period and for this starting concentration, shows a linear dependence (R2 = 0.996).

Figure 3: Real time measurement of 10 µM Pb2+ electrodeposition at the BDD electrode using in situ EC-XRF. (a) Plot of Pb XRF intensity over the energy range 9.5 to 11.5 keV versus time, acquired every five minutes; (b) Plot of XRF PbLα peak area from the data in (a), versus time. Inset: chronoamperometric i-t curve for the electrodeposition carried out in (a) over 60 mins. The effect of concentration on the EC-XRF signal was investigated for five concentrations of [Pb2+] over the range 1 – 10 µM, in order to determine the suitability of insitu EC-XRF for quantitative measurements. Data were collected every 50 s, for tdep = 60 12 ACS Paragon Plus Environment

Page 12 of 22

Page 13 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

mins, Edep = -1.5V vs. Ag/AgCl and Vf = 14 mL min-1. Figure 4a shows XRF intensity versus energy spectra, over the range 10.1 - 11 keV, for the different concentrations investigated, at the longest deposition time i.e. tdep = 60 min, where the signal is largest (as shown in Figure 3a). For each, the PbLσ peak is integrated and plotted versus concentration, as shown in Figure 4b, where a strong correlation between concentration and peak area exists (R2 = 0.994). From this data, the limit of detection, LOD, (3σ) was determined as 99 nM (20.5 ppb), close to the EPA limit for drinking water (75 nM; 15 ppb).35 Note, electrochemical preconcentration was recently combined with laser-induced breakdown spectroscopy, to detect Zn on Cu electrodes in-situ, however detection limits of only ppm could be achieved.36

Figure 4: (a) XRF intensity versus energy spectra for [Pb2+] solutions (in 0.2 M KNO3) over the concentration range 1 – 10 µM, tdep = 60 min and Edep = -1.5 V; (b) Plot of peak area versus concentration from the data in (a); (c) Plot of peak area versus tdep for the different concentrations investigated; (d) Plot of gradient in (c) vs [Pb2+].

13 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

From the measurements it is also possible to produce time-dependent data for each [Pb2+] concentration, as shown in Figure 4c. Each data point corresponds to an individual XRF spectrum, collected for 50 s and integrated to calculate the area under the PbLα line. From Figure 4c it can be clearly seen that the gradient of the peak area vs. t plot, which is itself a measurement of the deposition rate, increases as [Pb2+] increases, with a correlation coefficient, R2 = 0.991. As indicated by Figures 3b and 4c, and our previous measurements using ex-situ EC-XRF,24 depositing for longer, increasing the spectral acquisition time or further increasing mass transport, are some of the ways to increase detection sensitivity using this methodology. Finally, one of the drivers of the development of in situ EC-XRF is as a means to identify - and ultimately quantify - solutions containing multiple metals directly at the source, where the heavy metal composition of the mixture is unknown. In initial preliminary studies aimed at investigating the ability of in-situ EC-XRF to detect multiple metals, a solution containing six metals Fe3+, Ni2+, Cu2+, Zn2+, Hg2+ and Pb2+ at concentrations of 10 µM each, with 0.2 M KNO3 as supporting electrolyte, was prepared (pH 4.6). To illustrate the power of in-situ EC-XRF, Figure 5a shows the EC-only square wave (SW) stripping voltammetric response (4 mV steps, 10 mV amplitude, 25 Hz) recorded under stirred conditions (magnetic flea), with Edep = -1.5 V vs Ag/AgCl and tdep = 5 (red line) and 10 mins (black line), using a 1 mm diameter BDD macroelectrode. Edep was chosen to ensure a sufficiently high driving force for electrodeposition of all six metals. Evident in Figure 5a are five peaks, labelled (i) to (v), occurring at potentials of (i) -1.08 V; (ii) -0.48 V; (iii) -0.10 V; (iv) 0.26 V and (v) 0.87 V vs Ag/AgCl. Even with prior knowledge of the metals present in solution and consideration of their Eo values,37 as listed in Table 1, unambiguous chemical identification of these peaks is challenging. For example, based only on Eo values and the observed peak positions in Figure 5a, peaks (i) to (v) could be tentatively ascribed to (i) Zn;

14 ACS Paragon Plus Environment

Page 14 of 22

Page 15 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(ii) Ni or Pb; (iii) Fe; (iv) Cu and (v) Hg. However as Eo values do not account for the kinetics of electrodeposition, the effects of metal morphology,15 supporting electrolyte,38 alloying,39 electrocatalysis40 etc further experiments are required. These factors also account for the limited application of ASV in complex environmental systems.

Figure 5: (a) Square-wave ASV recorded in a stirred solution containing 10 µM of the following ions, Hg2+, Pb2+, Cu2+, Ni2+, Zn2+ and Fe3+ in 0.2 M KNO3 after Edep = -1.5 V and tdep = 5 mins (red line) and 10 mins (black line). (b) EC-XRF recorded in the same solution as (a), as a function of time for Edep = 1.5 V, tdep = 90 mins and Vf = 14 ml min-1. XRF spectra were recorded every 55 s. XRF enables unambiguous identification of the different elements present on the surface of the electrode. (c) XRF-only spectrum (red-line) of the six metals present at 1 mM concentration, evaporatively dried onto an XRF “Ultra-CarryTM” substrate. EC-XRF spectra (black line) taken from (b) at tdep = 90 mins. Note for both the data has been normalized with respect to the maximum Pb2+ intensity for respective spectra, to enable a useful comparison. (d) From (c), plot of the peak area versus deposition time for each metal. The data has been normalized with respect to the maximum peak area for each metal to enable comparison. Table 1: (1) Eo values for the electrochemical reduction of all six metals corrected against the Ag/AgCl reference electrode, (2) experimentally measured SW-ASV Epeak in individual metal salt solutions, and (3) energies of the strongest XRF emission peak.

15 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Metal Hg2+ Cu2+ Fe3+ Pb2+ Ni2+ Zn2+

(1) Eo vs NHE corrected against Ag/AgCl37 0.60 0.14 -0.24 -0.33 -0.53 -0.96

Page 16 of 22

(2) Epeak/ V vs Ag/AgCl

(3) XRF strongest peak intensities (keV)

0.18 -0.17 and 0.05 n/a -0.49 -0.48 and 0.82 -1.19

9.99 (Lα) 8.05 (Kα) 6.41 (Kα) 10.56 (Lα) 7.48 (Kα) 8.64 (Kα)

Individual SW-ASVs were thus recorded in six individual metal solutions under stirred conditions (10 µM of the metal salt and 0.2 M KNO3) using the same electrode and SW-ASV measurement conditions as for Figure 5a with tdep = 5 mins, ESI, section 3. From this data, potentials corresponding to the maximum SW stripping peak currents, Epeak were obtained as listed in Table 1. Points of note, from ESI, Figure S3, (i) no evidence of Fe electrodeposition/stripping was observed; (ii) Cu displayed two stripping peaks; (iii) the Ni peak at -0.48 V occurred at a similar potential to Pb, but the Pb stripping peak was significantly larger than for Ni and (iv) there was a second very small Ni stripping peak at 0.82 V (observed more readily after tdep = 10 mins). The origin of this second peak is difficult to ascertain; it may be due to oxidation of Ni2+ (Ni(OH)2) to Ni3+ (NiOOH) and is small due to the slightly acidic pH conditions of the solution.41 From these additional experiments we can now better assign the five SW-ASV peaks to (i) Zn; (ii) Pb and Ni; (iii) Cu; (iv) Hg and (v) Ni. In contrast using in-situ EC-XRF unambiguous identification of the metals present in solution is possible. Figure 5b shows the in-situ EC-XRF time-dependent response under experimental conditions of Vf = 14 mL min-1, Edep = -1.5 V and tdep = 90 mins, the latter employed to ensure sufficient deposition resulting in a detectable signal. Note the spectra in Figure 5 have been background subtracted due to the high Zn background (Figure 2a) vide supra. With knowledge of the characteristic metal XRF fluorescence lines, also given in 16 ACS Paragon Plus Environment

Page 17 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Table 1, it is now unequivocally clear which metals have electroplated onto the BDD electrode (Zn, Pb, Ni, Cu and Hg) and are thus present in the measurement solution. The absence of Fe (no peak was seen at 6.41 keV) correlates with the EC-only data. However, this is not surprising as the literature reports electrodeposition of Fe3+ is most favored under strongly acidic conditions.42 To understand the significance of the peak heights in the EC-XRF spectra, for comparison an XRF-only spectrum (signal acquisition time = 5 mins) of a solution containing the six metals at a concentration of 1 mM each (to ensure a signal would be seen), evaporatively dried (once) onto an UltracarryTM XRF substrate (Figure 5c) was recorded (red line). The intensities have been normalized with respect to that for Pb2+ (Pb2+ and Hg2+ gave the highest, very similar, peak intensities). The difference in peak intensities for the different metals, all present at the same concentration, reflects the different XRF efficiencies of individual elements. Fe3+ is clearly present in this spectrum. Also plotted in Figure 5c is the EC-XRF data (black line) from Figure 5b recorded at tdep = 90 mins, also normalized with respect to the highest intensity Pb signal. Qualitatively the EC-XRF signal intensities follow the same trend as for the XRF-only measurements, for all but Hg, boding well for calibration and quantification in mixed metal solutions. Less Hg than expected was deposited on the BDD electrode and could be due to the flow profile of the wall jet washing away some of the electroplated liquid Hg. When the time-dependent electrodeposition EC-XRF data, plotted as a function of peak area versus time, is displayed, normalized by the maximum peak area for each metal to enable comparison, the gradients appear to overlap (range: 0.00959 – 0.0128 min-1; 0.0114 ± .0013 min-1; average ± 1 standard deviation). The gradients represent the deposition rate of the different metals on the BDD surface and suggest that at the large overpotential employed (for the concentration stated) the deposition rates of all five metals are similar. This also

17 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

suggests that quantitative analysis in complex mixed metal solutions, using EC-XRF is possible, but is beyond the scope of this present work. CONCLUSIONS In this paper we introduce in-situ EC-XRF as a new methodology to both identify and ultimately quantify heavy metals directly in solution. The EC component employs a freestanding BDD electrode biased at a suitable overpotential to electrodeposit metals from solution onto the electrode. The X-rays pass through the back face of the BDD electrode and are used to directly interrogate the time-dependent XRF profile of the metals as they plate onto the electrode surface. In this way it is both possible to uniquely identify which metals are present in solution (by virtue of their characteristic XRF signature after preconcentration by electrodeposition), monitor metal deposition rates and ultimately quantify the concentration of metal ions in solution. Solution is flown over the BDD surface using a wall jet configuration, to enhance mass transport of species to the electrode during deposition. For the direct detection of Pb2+ in solution, XRF-only measurement results in a LOD of 0.682 mM, however by switching to EC preconcentration strategies the LOD is reduced to 99 nM (in this present configuration) i.e. by over three orders of magnitude. LODs could be reduced further by e.g. moving to thinner free-standing BDD electrodes, increasing mass transport, increasing pre-concentration and XRF signal collection times etc. Proof-of-concept multi metal detection was demonstrated using EC-XRF to detect the presence of six metals, Hg2+, Pb2+, Cu2+, Ni2+, Zn2+ and Fe3+ (all at 10 µM concentration in 0.2 M KNO3) in one solution. Measurements were compared to SW-ASV analysis in the same solution. EC-XRF was able to unequivocally identify five metals in one single measurement. In contrast, ASV required significant further experiments to enable identification of the five peaks evident in the SW-ASV trace. Both studies demonstrated that Fe3+ did not directly reduce to Fe under the experimental pH conditions employed. Note, we 18 ACS Paragon Plus Environment

Page 18 of 22

Page 19 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

have shown previously that simple electrochemical strategies can be employed to optimize the pH of the measurement environment in order to facilitate the electrochemical measurement of interest43,44 Similar strategies could be adopted here, when required. The relative intensity of the EC-XRF signals correlated with those recorded using XRFonly (evaporated samples) for all but Hg, suggesting less Hg than expected had plated onto the electrode surface. This is most likely due to some of the liquid metal being washed from the surface during deposition. In-situ EC-XRF also enabled the deposition rate of all five metals electroplating on the electrode surface to be monitored. Under the experimental conditions employed all five showed similar rates, boding well for quantitative analysis. Experiments are now underway to lower detection limits and provide a quantitative understanding of EC-XRF detection in real, multi-metal solutions. These are with a view to producing custom designed portable EC-XRF instrumentation to make sub-pbb heavy metal identification and analysis at the source a realistic possibility.

ACKNOWLEGEMENTS The authors thank Element Six for funding of GDO and for providing the freestanding BDD electrodes used throughout, Dr. Jonathan Newland for the photograph in Figure 1c and laser machining the BDD counter electrode, Mr. Ronald Wesson and Mr. Marcus Grant for assistance in fabricating the devices used herein and Mr. Lingcong Meng for collecting the SW-ASV data.

ASSOCIATED CONTENT Supporting Information. (1) Optimization of Electrode Deposition Parameters for Pb2+; (2) The effect of the rate of spectral acquisition on XRF response; (3) Interpretation of the square

19 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

wave-anodic stripping (SW-ASV) response of an aqueous solution containing six different dissolved metals. This material is available free of charge via the Internet at http://pubs.acs.org. Table of Contents Graphic and Text

In-situ EC-XRF: Solution containing trace metals is flown over a BDD electrode, where both electrochemical preconcentration and metal identification and quantification by XRF take place

REFERENCES (1)

Jarup, L. Br. Med. Bull. 2003, 68, 167–182.

(2)

Vallee, B. L.; Ulmer, D. D. Ann. Rev. Biochem. 1972, 41, 91–128.

(3)

Hamilton, J. W.; Kaltreider, R. C.; Bajenova, O. V; Ihnat, M. a; McCaffrey, J.; Turpie, B. W.; Rowell, E. E.; Oh, J.; Nemeth, M. J.; Pesce, C. a; Lariviere, J. P. Environ. Heal. Persp. 1998, 106, 1005–1015.

(4)

U.S. Environmental Protection Agency. Risk Assessment, Manage- ment and Communication of Drinking Water Contamination; US EPA 625/4-89/024, EPA: Washington, DC, 1989.

(5)

Off. J. Eur. Union L 364 2006, 5–24.

(6)

Aragay, G.; Pons, J.; Merkoçi, A. Chem. Rev. 2011, 111, 3433–3458.

(7)

Bersier, P. M.; Howell, J.; Bruntlett, C. Analyst 1994, 1, 219–232.

(8)

Ariza-Avidad, M.; Salinas-Castillo, A.; Cuéllar, M. P.; Agudo-Acemel, M.; Pegalajar, M. C.; Capitán-Vallvey, L. F. Anal. Chem. 2014.

20 ACS Paragon Plus Environment

Page 20 of 22

Page 21 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(9)

Liu, L.; Lin, H. Anal. Chem. 2014, 86, 8829–8834.

(10)

Herdan, J.; Feeney, R.; Kounaves, S. P.; Storment, C. W.; Kovacs, G. T. A.; Darling, R. B. Environ. Sci. Technol. 1998, 32, 131–136.

(11)

Feeney, R.; Kounaves, S. P. Anal. Chem. 2000, 72, 2222–2228.

(12)

Chumbimuni-Torres, K. Y.; Rubinova, N.; Radu, A.; Kubota, L. T.; Bakker, E. Anal. Chem. 2006, 78, 1318–1322.

(13)

De Marco, R.; Clarke, G.; Pejcic, B. Electroanalysis 2007, 19, 1987–2001.

(14)

Jones, S. E. W.; Chevallier, G.; Paddon, C. A.; Compton, R. G. Anal. Chem. 2007, 79, 4110–4119.

(15)

Hutton, L. A.; Newton, M. E.; Unwin, P. R.; Macpherson, J. V. Anal. Chem. 2011, 83, 735–745.

(16)

Bakker, E.; Bühlmann, P.; Pretsch, E. Chem. Rev. 1997, 97, 3083–3132.

(17)

Bernick, M. B.; Kalnicky, D. J.; Prince, G.; Singhvi, R. J. Hazard. Mater. 1995, 43, 101–110.

(18)

Melquiades, F. L.; Appoloni, C. R. J. Radioanal. Nucl. Chem. 2004, 262, 533–541.

(19)

Jang, M. Environ. Geochem. Health 2010, 32, 207–216.

(20)

Prange, A. Spectrochim. Acta Part B 1989, 44, 437–452.

(21)

Barreiros, M. A.; Carvalho, M. L.; Costa, M. M.; Marques, M. I.; Ramos, M. T. X-Ray Spectrom. 1997, 26, 165–168.

(22)

Peng, S.; Liu, Z.; Sun, T.; Ma, Y.; Ding, X. Anal. Chem. 2014, 86, 362–366.

(23)

Melquiades, F. L.; Parreira, P. S.; Appoloni, C. R.; Silva, W. D.; Lopes, F. Appl. Radiat. Isot. 2011, 69, 327–333.

(24)

Hutton, L. A.; O’Neil, G. D.; Read, T. L.; Ayres, Z. J.; Newton, M. E.; Macpherson, J. V. Anal. Chem. 2014, 86, 4566–4572.

(25)

Macpherson, J. V. Phys. Chem. Chem. Phys. 2015, 17, 2935–2949.

(26)

Hutton, L. A.; Iacobini, J. G.; Bitziou, E.; Channon, R. B.; Newton, M. E.; Macpherson, J. V. Anal. Chem. 2013, 85, 7230–7240.

(27)

Prado, C.; Wilkins, S. J.; Grundler, P.; Marken, F.; Compton, R. G. Electroanalysis 2003, 15, 1011–1016.

(28)

Hutton, L.; Newton, M. E.; Unwin, P. R.; Macpherson, J. V. Anal. Chem. 2009, 81, 1023–1032. 21 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(29)

Soucaze-Guillous, B.; Kutner, W. Electroanalysis 1997, 9, 32–39.

(30)

Bitziou, E.; Rudd, N. C.; Edwards, M. A; Unwin, P. R. Anal. Chem. 2006, 78, 1435– 1443.

(31)

Joseph, M. B.; Bitziou, E.; Read, T. L.; Meng, L.; Palmer, N. L.; Mollart, T. P.; Newton, M. E.; Macpherson, J. V. Anal. Chem. 2014, 86, 5238–5244.

(32)

Chin, T.; Tsang, H. J. Electrochem. Soc. 1978, 125, 1461–1470.

(33)

Brett, C. M. A.; Oliveira Brett, A. M. C. F.; Fisher, A. C.; Compton, R. G. J. Electroanal. Chem. 1992, 334, 57–64.

(34)

Macpherson, J. V; O’Hare, D.; Unwin, P. R.; Winlove, C. P. Biophys. J. 1997, 73, 2771–2781.

(35)

U.S. Environmental Protection Agency. National Primary Drinking Water Regulations. http://water.epa.gov/drink/contaminants.

(36)

Matsumoto, A.; Tamura, A.; Koda, R.; Fukami, K.; Ogata, Y. H.; Nishi, N.; Thornton, B.; Sakka, T. Anal. Chem. 2015.

(37)

Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; 2nd Ed.; Wiley: New York, 2001.

(38)

Saciloto, T. R.; Cervini, P.; Cavalheiro, É. T. G. Electroanalysis 2014, 26, 2664–2676.

(39)

Babyak, C.; Smart, R. B. Electroanalysis 2004, 16, 175–182.

(40)

Prado, C.; Wilkins, S. J.; Marken, F.; Compton, R. G. Electroanalysis 2002, 14, 262– 272.

(41)

Bossu, F. P.; Margerum, D. W. Inorg. Chem. 1977, 16, 1210–1214.

(42)

Izaki, M. In Modern Electroplating; Schlesinger, M.; Paunovic, M., Eds.; John Wiley & Sons, Inc.: Hoboken, NJ, 2010; pp. 309–326.

(43)

Read, T. L.; Bitziou, E.; Joseph, M. B.; Macpherson, J. V. Anal. Chem. 2014, 86, 367– 371.

(44)

Bitziou, E.; Joseph, M. B.; Read, T. L.; Palmer, N.; Mollart, T.; Newton, M. E.; Macpherson, J. V. Anal. Chem. 2014, 86, 10834–10840.

22 ACS Paragon Plus Environment

Page 22 of 22